Techniques are described for enabling users of a cloud provider network to launch compute instances with the option to dedicate networking paths between the compute instances in an otherwise shared networking environment. A user can send, for example, a request to launch two or more compute instances, where the request further indicates a request for a dedicated networking path between the instances. A hardware virtualization service launches the requested compute instances and further sends signals to the computing devices at which the instances are launched indicating the request for a dedicated networking path. A computing device, upon receiving such a signal for a hosted instance, configures routing information at the computing device to dedicate, to the identified instance, an available network link accessible to the computing device and further propagates the signal to additional networking devices as needed to form a dedicated networking path between the associated compute instances.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer-implemented method comprising: receiving, by a hardware virtualization service of a cloud provider network, a first request to launch a plurality of compute instances including a first compute instance and a second compute instance, wherein the request further requests a dedicated networking path between the first compute instance and the second compute instance; launching the first compute instance at a first computing device, wherein the first computing device is associated with a first plurality of network links to a first networking device, and wherein the first plurality of network links are shared by first compute instances running on the first computing device; launching the second compute instance at a second computing device, wherein the second computing device is associated with a second plurality of network links to a second networking device, and wherein the second plurality of network links are shared by second compute instances running on the second computing device; sending, to the first computing device, a second request to dedicate, to the first compute instance, a first network link of the first plurality of network links, wherein the second request causes the first computing device to select a first particular network link of the first plurality of network links to dedicate to the first compute instance, and wherein the second request further causes the first computing device to send a third request that causes one or more first networking devices including the first networking device to dedicate network links to the first compute instance; and sending a fourth request to the second computing device to dedicate, to the second compute instance, a second network link of the second plurality of network links, wherein the fourth request causes the second computing device to select a second particular network link of the second plurality of network links to dedicate to the second compute instance, and wherein the fourth request causes the second computing device to send a fifth request that causes one or more second networking devices including the second networking device to dedicate network links to the second compute instance.
This invention relates to cloud computing and network virtualization, specifically addressing the need for dedicated, low-latency networking paths between compute instances in a cloud provider network. In cloud environments, compute instances often share network links, leading to performance variability and potential congestion. The invention provides a method to establish a dedicated networking path between two compute instances, ensuring predictable performance and reduced latency. The method involves receiving a request to launch multiple compute instances with a requirement for a dedicated networking path between them. The first compute instance is launched on a first computing device connected to a first networking device via multiple shared network links. Similarly, the second compute instance is launched on a second computing device connected to a second networking device via multiple shared network links. To establish the dedicated path, the first computing device is instructed to select and dedicate a specific network link to the first compute instance, while the second computing device does the same for the second compute instance. These devices then communicate with their respective networking devices to ensure the network links are exclusively allocated to the compute instances, creating a direct, high-performance connection between them. This approach minimizes interference from other traffic, improving reliability and performance for applications requiring low-latency communication.
2. The computer-implemented method of claim 1 , wherein the third request causes the first networking device to dedicate the first particular network link by removing a route associated with the first compute instance from an equal-cost multi-path (ECMP) set specified in a routing table of the first networking device.
This invention relates to network traffic management in distributed computing environments, specifically addressing the challenge of dynamically allocating and deallocating network links to optimize performance and resource utilization. The method involves a system where a first networking device manages network links between compute instances, such as virtual machines or containers, to ensure efficient data transmission. The method includes a step where a third request is processed by the first networking device. This request triggers the device to dedicate a first particular network link by removing a route associated with a first compute instance from an equal-cost multi-path (ECMP) set. The ECMP set is part of the routing table of the first networking device and typically includes multiple paths of equal cost for load balancing traffic. By removing the route, the first compute instance is no longer part of the ECMP set, effectively dedicating the first network link exclusively to the first compute instance. This ensures that traffic for the first compute instance is routed through the dedicated link, improving performance and reducing latency for critical workloads. The method may also involve similar steps for other compute instances and network links, allowing dynamic reconfiguration of network paths based on demand or priority.
3. The computer-implemented method of claim 1 , wherein the third request further causes the first networking device to send a fifth request to an aggregation layer switch to dedicate a third network link of a third plurality of network links available at the aggregation layer switch to the first compute instance, and wherein the aggregation layer switch dedicates the third network link by removing a route associated with the first compute instance from an equal-cost multi-path (ECMP) routing table at the aggregation layer switch.
This invention relates to network traffic management in data centers, specifically optimizing network link allocation for compute instances. The problem addressed is inefficient use of network resources when multiple compute instances share bandwidth through equal-cost multi-path (ECMP) routing, leading to congestion and suboptimal performance. The method involves dynamically dedicating specific network links to individual compute instances to improve traffic flow. When a compute instance requires enhanced network performance, a request is sent to a networking device, which then initiates a series of actions to allocate a dedicated network link. This includes sending a request to an aggregation layer switch to dedicate a specific link from its available pool. The aggregation layer switch removes the compute instance's route from its ECMP routing table, effectively reserving the link exclusively for that instance. This ensures the compute instance receives prioritized bandwidth, reducing contention and improving network efficiency. The method supports scalable and flexible network resource allocation, adapting to varying workload demands while maintaining high performance for critical applications.
4. A computer-implemented method comprising: receiving, by a hardware virtualization service of a cloud provider network, a request to launch a plurality of compute instances including a first compute instance and a second compute instance, wherein the request further requests a dedicated networking path between the first compute instance and the second compute instance, wherein the dedicated networking path involves at least one intermediary networking device; launching the first compute instance at a first computing device and the second compute instance at a second computing device; and sending network link dedication requests to the first computing device and the second computing device, wherein the network link dedication requests cause the first computing device to dedicate a first network link to the first compute instance and cause the second computing device to dedicate a second network link to the second compute instance, and wherein the dedicated networking path between the first compute instance and the second compute instance includes the first network link and the second network link.
This invention relates to cloud computing and network virtualization, specifically addressing the need for dedicated, high-performance networking paths between compute instances in a cloud provider network. The problem solved is the lack of guaranteed low-latency, high-throughput connections between virtual machines (compute instances) in shared cloud environments, where network performance can be unpredictable due to shared infrastructure. The method involves a hardware virtualization service in a cloud provider network receiving a request to launch multiple compute instances (e.g., virtual machines) with a requirement for a dedicated networking path between them. This path includes at least one intermediary networking device, such as a switch or router. Upon receiving the request, the service launches the first compute instance on a first physical computing device and the second compute instance on a second physical computing device. The service then sends network link dedication requests to both computing devices, instructing them to reserve and dedicate specific network links (e.g., physical or virtual network interfaces) exclusively to their respective compute instances. The dedicated networking path is established by combining these reserved links, ensuring a direct, high-performance connection between the two compute instances while bypassing shared network resources that could introduce latency or congestion. This approach improves network reliability and performance for applications requiring low-latency communication, such as high-frequency trading or real-time data processing.
5. The computer-implemented method of claim 4 , wherein the network link dedication requests cause the first computing device to send, to a networking device of the at least one intermediary networking device, an additional network link dedication request, and wherein the additional network link dedication request causes the networking device to dedicate a network link to the first compute instance by removing a route associated with the first compute instance from an equal-cost multi-path (ECMP) set specified in a routing table of the networking device.
This invention relates to network traffic management in distributed computing environments, specifically addressing the problem of efficiently dedicating network links to compute instances to ensure predictable performance. In such environments, multiple compute instances may share network resources, leading to congestion and unpredictable latency. The invention provides a method to dynamically dedicate network links to specific compute instances by modifying routing tables in intermediary networking devices. The method involves sending network link dedication requests from a first computing device to at least one intermediary networking device. These requests trigger the networking device to dedicate a network link to a first compute instance by removing a route associated with that instance from an equal-cost multi-path (ECMP) set in the routing table. ECMP sets distribute traffic across multiple paths to balance load, but this can lead to contention. By removing the route from the ECMP set, the networking device ensures that traffic for the compute instance is routed exclusively through a dedicated link, reducing latency and improving performance. The method may also involve sending additional network link dedication requests to other networking devices in the path, ensuring end-to-end link dedication. This approach allows for dynamic allocation of network resources based on demand, improving efficiency and performance in distributed computing environments. The invention is particularly useful in cloud computing and data center networks where resource allocation must be flexible and scalable.
6. The computer-implemented method of claim 4 , wherein the network link dedication requests cause the first computing device to send, to a networking device of the at least one intermediary networking device, a first additional network link dedication request, wherein the first additional network link dedication request causes the networking device to send a second additional network link dedication request to an aggregation layer switch to dedicate a network link to the first compute instance, and wherein the aggregation layer switch dedicates the network link by removing a route associated with the first compute instance from an equal-cost multi-path (ECMP) routing table at the aggregation layer switch.
This invention relates to network traffic management in data center environments, specifically addressing the challenge of dynamically dedicating network links to compute instances to ensure low-latency, high-priority communication paths. In modern data centers, multiple compute instances often share network resources, leading to congestion and unpredictable latency. The invention provides a method to prioritize network traffic for specific compute instances by dedicating network links, thereby improving performance for critical workloads. The method involves a first computing device sending network link dedication requests to intermediary networking devices. These requests trigger a cascading process where a networking device forwards a first additional network link dedication request to an aggregation layer switch. The aggregation layer switch then dedicates a network link to the first compute instance by removing its route from an equal-cost multi-path (ECMP) routing table. This removal ensures the compute instance's traffic is no longer subject to ECMP load balancing, guaranteeing a dedicated path. The method may also involve similar requests for other compute instances, allowing dynamic allocation of network resources based on workload demands. This approach enhances network efficiency by reducing contention and ensuring predictable performance for high-priority traffic.
7. The computer-implemented method of claim 4 , wherein the hardware virtualization service includes a data store storing state information about network links managed by the hardware virtualization service, and wherein the data store indicates, for each of a plurality of network links, a pair of devices connected by the network link, whether the network link is part of a shared pool of network links or dedicated to a compute instance, and performance characteristics of the network link; and wherein the method further comprises updating the data store to reflect the dedication of network links for the dedicated networking path.
This invention relates to hardware virtualization services that manage network links between compute instances. The problem addressed is the need to efficiently allocate and track network links in virtualized environments, particularly when dedicating links for high-performance or secure networking paths. The system includes a data store that maintains state information about network links managed by the hardware virtualization service. For each network link, the data store records the pair of devices connected by the link, whether the link is part of a shared pool or dedicated to a specific compute instance, and performance characteristics of the link. This allows the service to dynamically allocate and track network resources. When a dedicated networking path is requested, the method updates the data store to reflect the dedication of the network links for that path. This ensures that the links are reserved and their performance characteristics are properly accounted for, preventing conflicts with other network allocations. The system enables efficient management of network resources in virtualized environments, supporting both shared and dedicated networking configurations.
8. The computer-implemented method of claim 4 , wherein the hardware virtualization service includes a data store storing state information about network links managed by the hardware virtualization service, wherein the state information about the network links includes network capacity information associated with the network links and with a networked computing environment in which the first compute instance and second compute instance operate, and wherein the method further comprises determining, based on state information stored in the data store, whether sufficient networking capacity exists to dedicate network links for the dedicated networking path.
This invention relates to hardware virtualization services in networked computing environments, specifically addressing the challenge of ensuring sufficient network capacity for dedicated networking paths between compute instances. The method involves a hardware virtualization service that manages network links and maintains a data store containing state information about these links. The state information includes network capacity details, such as bandwidth availability, for both individual network links and the broader networked computing environment where the compute instances operate. The method further includes determining whether there is enough networking capacity to establish a dedicated networking path between a first compute instance and a second compute instance by analyzing the stored state information. This ensures that the dedicated path can be provisioned without overloading the network infrastructure. The solution enhances network performance and reliability by dynamically assessing and allocating network resources based on real-time capacity data.
9. The computer-implemented method of claim 4 , further comprising: receiving a request to release network links of the dedicated networking path back into one or more shared pools of network links; and sending dedicated network link release requests to the first computing device and to the second computing device, wherein the dedicated network link release requests cause the first computing device to release the first network link to a first shared pool of network links associated with the first computing device and further cause the second computing device to release the second network link to a second shared pool of network links associated with the second computing device.
This invention relates to network resource management in computing systems, specifically the dynamic allocation and release of dedicated network links between computing devices. The problem addressed is the inefficient use of network resources when dedicated paths are established for specific tasks but remain unused after completion, leading to underutilized network capacity. The method involves managing network links between a first computing device and a second computing device. Initially, a dedicated networking path is established by allocating a first network link from a first shared pool associated with the first computing device and a second network link from a second shared pool associated with the second computing device. This dedicated path ensures high-performance, low-latency communication for specific tasks. When the dedicated path is no longer needed, a request is received to release the network links back into their respective shared pools. The system then sends dedicated network link release requests to both computing devices. Upon receiving these requests, the first computing device releases the first network link back to its shared pool, and the second computing device releases the second network link back to its shared pool. This process optimizes network resource utilization by making previously dedicated links available for other tasks, improving overall network efficiency.
10. The computer-implemented method of claim 4 , wherein the first compute instance and the second compute instance implement at least a portion of a distributed workload, and wherein execution of the distributed workload involves the exchange of network traffic between the first compute instance and the second compute instance using the dedicated networking path.
This invention relates to distributed computing systems where multiple compute instances collaborate to execute a workload, requiring efficient and secure communication between them. The problem addressed is ensuring reliable, low-latency, and secure data exchange between compute instances in a distributed workload environment, particularly in cloud or virtualized systems where network performance and security are critical. The method involves establishing a dedicated networking path between a first compute instance and a second compute instance. This dedicated path is used to exchange network traffic generated during the execution of a distributed workload. The workload is divided such that at least a portion of it is processed by each compute instance, necessitating inter-instance communication. The dedicated networking path ensures that traffic related to the workload is prioritized, isolated, or otherwise optimized compared to general network traffic, improving performance and security. This approach is particularly useful in scenarios where workloads involve real-time processing, sensitive data, or strict latency requirements. The method may include configuring network policies, allocating bandwidth, or encrypting traffic to enhance security and efficiency. By using a dedicated path, the system avoids congestion and interference from unrelated network traffic, ensuring consistent and predictable performance for the distributed workload.
11. The computer-implemented method of claim 4 , wherein the request identifies a processing cluster into which the compute instances are to be launched, and wherein the hardware virtualization service places the first compute instance and the second compute instance at the first computing device and the second computing device according to a placement strategy associated with the processing cluster.
This invention relates to a computer-implemented method for managing the deployment of compute instances in a cloud computing environment. The method addresses the challenge of efficiently distributing compute instances across physical computing devices to optimize resource utilization and performance. When a request is made to launch compute instances, the system identifies a specific processing cluster where the instances should be deployed. The hardware virtualization service then places the first compute instance on a first computing device and the second compute instance on a second computing device based on a predefined placement strategy associated with the processing cluster. The placement strategy ensures that the compute instances are distributed in a manner that balances workload, minimizes latency, or adheres to other operational constraints. The method may also involve dynamically adjusting the placement of instances within the cluster to maintain optimal performance as demand fluctuates. This approach improves resource allocation, reduces overhead, and enhances the scalability of cloud computing services.
12. The computer-implemented method of claim 4 , further comprising: receiving a request to reserve a dedicated networking path for two or more compute instances; and storing a dedicated network path reservation including data representing a set of network links to be used to establish the dedicated networking path upon request.
This invention relates to network resource management in cloud computing environments, specifically addressing the challenge of ensuring reliable, low-latency communication between compute instances by reserving dedicated networking paths. The method involves receiving a request to reserve a dedicated networking path for two or more compute instances, which may be virtual machines or other cloud-based processing units. The system then stores a reservation that includes data representing a set of network links to be used to establish the dedicated path when requested. This reservation ensures that the specified network links are reserved and available for use, preventing interference from other traffic and guaranteeing consistent performance. The method may also include dynamically configuring the reserved path based on network conditions or traffic requirements, ensuring optimal performance for the connected compute instances. By pre-allocating network resources, this approach minimizes latency and packet loss, which is critical for applications requiring high reliability, such as real-time data processing or distributed computing tasks. The invention improves upon traditional cloud networking by providing predictable, high-performance connectivity between compute instances.
13. The computer-implemented method of claim 4 , wherein the at least one intermediary networking device is part of a networking environment including a hierarchy of networking devices connecting computing devices of one or more data centers in which the hardware virtualization service provisions compute instances.
This invention relates to network management in data center environments, specifically addressing challenges in monitoring and controlling traffic flow through intermediary networking devices in a hierarchical network architecture. The system involves a hardware virtualization service that provisions compute instances across one or more data centers, where these instances communicate through a network hierarchy comprising multiple intermediary networking devices. The method focuses on dynamically managing traffic routing and performance optimization within this environment. The intermediary networking devices, which may include switches, routers, or gateways, are configured to handle data packets exchanged between compute instances and external networks. The system monitors traffic patterns, identifies congestion points, and adjusts routing paths to maintain efficient data flow. Additionally, the method may involve load balancing, quality-of-service (QoS) enforcement, and security policy application at the intermediary devices to ensure reliable and secure communication. The hierarchical structure allows for scalable management of network resources, enabling efficient provisioning and deprovisioning of compute instances while maintaining optimal network performance. The invention aims to improve network reliability, reduce latency, and enhance overall data center operational efficiency by intelligently managing traffic through the intermediary devices in the network hierarchy.
14. The computer-implemented method of claim 4 , wherein the network link dedication requests are sent to the first computing device and to the second computing device using network link dedicated to control signal information sent and received by the first computing device and the second computing device.
This invention relates to network communication systems where dedicated network links are used for control signal information between computing devices. The problem addressed is the need for efficient and reliable transmission of control signals in networked environments, particularly where multiple devices must coordinate their operations. The invention provides a method for sending network link dedication requests over a dedicated control signal link, ensuring that control information is prioritized and isolated from other data traffic. The method involves establishing a dedicated network link specifically for control signal information between a first computing device and a second computing device. When a network link dedication request is generated, it is transmitted over this dedicated link rather than a general-purpose data link. This ensures that control signals are not delayed or disrupted by other network traffic, improving system reliability and performance. The dedicated link may be used for both sending and receiving control signals, allowing bidirectional communication. The method is particularly useful in systems where timely control signal transmission is critical, such as in industrial automation, telecommunication networks, or distributed computing environments. By separating control signals from general data traffic, the invention reduces latency and interference, enhancing overall system coordination and stability.
15. The computer-implemented method of claim 4 , wherein the first computing device and the second computing device are located in a provider substrate extension that is separate from the cloud provider network.
This invention relates to distributed computing systems, specifically methods for managing computing resources in a provider substrate extension that operates independently from a cloud provider network. The technology addresses the challenge of efficiently coordinating and utilizing computing resources across multiple devices in a decentralized environment, particularly when these devices are physically or logically separated from the primary cloud infrastructure. The method involves a first computing device and a second computing device, both situated within a provider substrate extension that is distinct from the cloud provider network. The first computing device generates a request for a computing task, which is then transmitted to the second computing device. The second computing device processes this request and executes the task, generating a result that is subsequently transmitted back to the first computing device. This interaction ensures that computing tasks can be distributed and managed effectively within the provider substrate extension, even when operating outside the cloud provider network's direct control. The method may also include additional steps such as authentication, task prioritization, or resource allocation to optimize performance and security within the decentralized system. The approach enables scalable and flexible computing operations in environments where direct cloud integration is not feasible or desirable.
16. A system comprising: a first one or more electronic devices to implement a hardware virtualization service in a cloud provider network, the hardware virtualization service including instructions that upon execution cause the hardware virtualization service to: receive, a request to launch a plurality of compute instances including a first compute instance and a second compute instance, wherein the request further requests a dedicated networking path between the first compute instance and the second compute instance, wherein the dedicated networking path involves at least one intermediary networking device; launch the first compute instance at a first computing device and the second compute instance at a second computing device; and send network link dedication requests to the first computing device and the second computing device; and a second one or more electronic devices to implement the first computing device, the first computing device including instructions that upon execution cause the first computing device to: receive a network link dedication request identifying the first compute instance hosted by the first computing device; select a network link associated with the first computing device to dedicate to the first compute instance; modify networking configuration information to dedicate the network link to the first computing instance; and send, to a next-hop networking device, a request to dedicate a network link to the first compute instance.
This system relates to cloud computing and network virtualization, specifically addressing the need for dedicated networking paths between compute instances in a cloud provider network. The system enables users to request a dedicated networking path between multiple compute instances, ensuring low-latency, high-bandwidth, and secure communication channels. The system includes a hardware virtualization service that processes launch requests for compute instances, including requests for dedicated networking paths. Upon receiving such a request, the service launches the compute instances on separate computing devices and sends network link dedication requests to those devices. Each computing device then selects a network link to dedicate to its respective compute instance, modifies its networking configuration to enforce this dedication, and forwards a request to the next-hop networking device to extend the dedicated path. This ensures end-to-end network isolation and performance guarantees for the specified compute instances. The system automates the provisioning of dedicated network links, reducing manual configuration and improving scalability in cloud environments.
17. The system of claim 16 , wherein the network link dedication requests cause the first computing device to send, to a networking device, an additional network link dedication request, and wherein the additional network link dedication request causes the networking device to remove a route associated with the first compute instance from an equal-cost multi-path (ECMP) set specified in a routing table of the networking device.
This invention relates to network traffic management in computing systems, specifically addressing the challenge of dynamically adjusting network paths to optimize performance and resource allocation. The system involves a first computing device that manages network link dedication requests for compute instances, ensuring efficient routing of data traffic. When a network link dedication request is made, the first computing device sends an additional request to a networking device, instructing it to remove a route associated with a specific compute instance from an equal-cost multi-path (ECMP) set in the routing table. This adjustment prevents the compute instance from utilizing multiple redundant paths, thereby dedicating the network link exclusively to that instance. The system ensures that traffic for the compute instance is routed through a single, dedicated path, improving predictability and performance. The networking device updates its routing table accordingly, removing the route from the ECMP set to enforce the dedicated link. This approach is particularly useful in environments where consistent, low-latency communication is critical, such as in high-performance computing or real-time applications. The invention enhances network efficiency by dynamically managing path selection based on compute instance requirements.
18. The system of claim 16 , wherein the network link dedication requests cause the first computing device to send, to a networking device of the at least one intermediary networking device, a first additional network link dedication request, wherein the first additional network link causes the networking device to send a second additional network link dedication request to an aggregation layer switch to dedicate a network link to the first compute instance, and wherein the aggregation layer switch dedicates the network link by removing a route associated with the first compute instance from an equal-cost multi-path (ECMP) routing table at the aggregation layer switch.
This invention relates to network link dedication in a computing environment, specifically addressing the challenge of dynamically allocating and managing network links for compute instances to ensure low-latency, high-performance communication. The system involves a first computing device that generates network link dedication requests to establish dedicated network paths for a first compute instance. These requests trigger a cascading process where the first computing device sends a first additional network link dedication request to an intermediary networking device. The intermediary networking device then forwards a second additional network link dedication request to an aggregation layer switch. The aggregation layer switch responds by dedicating a network link to the first compute instance by removing a route associated with the compute instance from its equal-cost multi-path (ECMP) routing table. This removal ensures that traffic for the compute instance is routed through a dedicated link rather than being distributed across multiple paths, thereby reducing latency and improving performance. The system may also include a second computing device that sends a second network link dedication request to the first computing device, which then forwards it to the intermediary networking device to establish a dedicated link for a second compute instance. The aggregation layer switch similarly removes the corresponding route from its ECMP routing table to enforce the dedicated path. This approach optimizes network resource allocation by dynamically adjusting routing tables to prioritize critical traffic flows.
19. The system of claim 16 , wherein the hardware virtualization service includes a data store storing state information about network links managed by the hardware virtualization service, and wherein the data store indicates, for each of a plurality of network links, a pair of devices connected by the network link, whether the network link is part of a shared pool of network links or dedicated to a compute instance, and performance characteristics of the network link; and wherein the hardware virtualization service further includes instructions that upon execution cause the hardware virtualization service to update the data store to reflect the dedication of network links for the dedicated networking path.
A hardware virtualization system manages network links between compute instances and physical devices. The system addresses the challenge of efficiently allocating and tracking network resources in virtualized environments, ensuring optimal performance and isolation for dedicated networking paths. The system includes a data store that maintains state information about network links, including the devices connected by each link, whether the link is part of a shared pool or dedicated to a specific compute instance, and performance characteristics such as bandwidth and latency. The system dynamically updates this data store to reflect changes in network link allocation, particularly when links are dedicated for exclusive use by a compute instance. This allows the system to enforce network isolation, prioritize traffic, and optimize resource utilization based on real-time requirements. The solution enhances network management in virtualized environments by providing granular control over network link assignments and performance tracking.
20. The system of claim 16 , wherein the hardware virtualization service includes a data store storing state information about network links managed by the hardware virtualization service, wherein the state information about the network links includes network capacity information associated with the network links and with a networked computing environment in which the first compute instance and second compute instance operate, and wherein the hardware virtualization service further includes instructions that upon execution cause the hardware virtualization service to update determine, based on state information stored in the data store, whether sufficient networking capacity exists to dedicate network links for the dedicated networking path.
A hardware virtualization system manages network links between compute instances in a networked computing environment. The system addresses the challenge of ensuring reliable, high-performance networking between virtualized compute instances by dynamically assessing and allocating network capacity. The system includes a hardware virtualization service that maintains a data store containing state information about network links, including their capacity and the overall network environment. This state information allows the service to evaluate whether sufficient networking capacity exists to establish dedicated networking paths between compute instances. The system dynamically determines network availability and allocates resources to ensure optimal performance for critical workloads. By monitoring and managing network link states, the system prevents bottlenecks and ensures consistent connectivity for virtualized compute instances. The solution is particularly useful in cloud computing and virtualized environments where network performance directly impacts application efficiency. The system's ability to assess and allocate network capacity in real-time enhances reliability and scalability for distributed computing workloads.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
June 29, 2020
March 22, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.